Fundamentals of radio astronomy James Miller-Jones (NRAO Charlottesville/Curtin University)
ITN 215212: Black Hole Universe Many slides taken from NRAO Synthesis Imaging Workshop (http://www.aoc.nrao.edu/events/synthesis/2010/) and the radio astronomy course of J. Condon and S. Ransom Atacama Large Millimeter/submillimeter Array (http://www.cv.nrao.edu/course/astr534/ERA.shtml) Expanded Very Large Array Robert C. Byrd Green Bank Telescope Very Long Baseline Array What is radio astronomy?
• The study of radio waves originating from outside the Earth • Wavelength range 10 MHz – 1 THz • Long wavelengths, low frequencies, low photon energies – Low-energy transitions (21cm line) – Cold astronomical sources – Stimulated emission (masers) – Long-lived synchrotron emission • Sources typically powered by gravity rather than nuclear fusion – Radio galaxies – Supernovae – Pulsars – X-ray binaries
2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 2 What is radio astronomy?
• Telescope resolution q = 1.02 l/D – Requires huge dishes for angular resolution • Coherent (phase preserving) amplifiers practical
– Quantum noise proportional to frequency T = hn/kB • Precision telescopes can be built s 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 3 A short historical digression • Extraterrestrial radio signals first detected in 1932 by Karl Jansky – Electrical engineer working for Bell Labs – Attempting to determine origin of radio static at 20.5 MHz – Steady hiss modulated with a period 23h 56min – Strongest towards the Galactic Centre • Discovery ignored by most astronomers, pursued by Grote Reber – 1937-1940: Mapped the Galaxy at 160 MHz – Demonstrated the spectrum was non-thermal • World War II stimulated rapid progress in radio and radar technology 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 4 Why radio astronomy? • Atmosphere absorbs most of the EM spectrum – Only optical and radio bands accessible to ground-based instruments • Rayleigh scattering by atmospheric dust prevents daytime optical • Radio opacity from water vapour line, hydrosols, molecular oxygen – High, dry sites best at high frequencies • Stars undetectably faint as thermal radio sources – Non-thermal radiation – Cold gas emission – CMB 2hn 3 1 2k Tn 2 B T B 0.3Jy n c2 hn c2 e kBT 1 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 5 A tour of the radio Universe 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 6 What are we looking at? • Synchrotron radiation – Energetic charged particles accelerating along magnetic field lines (non-thermal) • What can we learn? • particle energy • strength of magnetic field • polarization • orientation of magnetic field 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 7 What are we looking at? • Thermal emission – Blackbody radiation from cool (T~3-30K) objects – Bremsstrahlung “free-free” radiation: charged particles interacting in a plasma at T; e- accelerated by ion • H II regions • What can we learn? • mass of ionized gas • optical depth • density of electrons in plasma • rate of ionizing photons 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 8 What are we looking at? • How to distinguish continuum mechanisms? • Look at the spectrum Spectral index across a SNR Flux or Flux Density Density Flux or Flux Frequency • Spectral index From T. Delaney Sn n 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 9 What are we looking at? • Spectral line emission (discrete transitions of atoms and molecules) – Spin-flip 21-cm line of HI – Recombination lines – Molecular rotational/vibrational modes • What can we learn? • gas physical conditions • kinematics (Doppler effect) 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 10 The Sun • Brightest discrete radio source • Atmospheric dust doesn‟t scatter radio waves Image courtesy of NRAO/AUI 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 11 The planets • Thermal emitters • No reflected solar radiation detected • Jupiter is a non-thermal emitter (B field traps electrons) • Also active experiments: radar Jupiter Mars (radar) Saturn Images courtesy of NRAO/AUI 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 12 Stars • Coronal loop resolved in the Algol system • Tidally-locked system, rapid rotation drives magnetic dynamo Credit: Peterson, Mutel, Guedel, Goss (2010, Nature) 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 13 Black holes and neutron stars Blundell & Bowler (2004) Mirabel & Rodriguez (1994) Dhawan et al. (2000) Dhawan et al. (2006) Gallo et al. (2005) Dubner et al. (1998) 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 14 Center of our Galaxy VLA 20cm VLA 1.3cm VLA 3.6cm 1 LY SgrA* - 4 milliion Mo black hole source15 Credits: Lang, Morris, Roberts, Yusef-Zadeh, Goss, Zhao Radio Spectral Lines: Gas in galaxies 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 16 Jet Energy via Radio Bubbles in Hot Cluster Gas 1.4 GHz VLA contours over Chandra X-ray image (left) and optical (right) 6 X 1061 ergs ~ 3 X 107 solar masses X c2 (McNamara et al. 2005, Nature, 433, 45) 17 Resolution and Surface-brightness Sensitivity 18 The Cosmic Microwave Background • 2.73 K radiation from z~1100 • Angular power spectrum of brightness fluctuations constrains cosmological parameters 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 19 Definitions 2 In (or Bn ) = Surface Brightness : erg/s/cm /Hz/sr (= intensity) S = Flux density : erg/s/cm 2/Hz I n n S = Flux : erg/s/cm 2 I n n P = Power received : erg/s I nA n tel E = Energy : erg I nA t n tel 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 20 Radio astronomy: instrumentation • Antennas convert electromagnetic radiation into electrical currents in conductors • Average collecting area of any lossless Prime focus Cassegrain antenna (GMRT) l2 (ATCA) A e 4 – Long wavelength: dipoles sufficient – Short wavelength: use reflectors to Offset cassegrain Naysmith collect and focus power from a large (EVLA) (OVRO) area onto a single feed • Parabola focuses a plane wave to a single focal point Dual offset Beam waveguide offset (NRO) (GBT) 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 21 Types of antennas • Wire antennas (l 1m) – Dipole – Yagi – Helix – Small arrays of the above • Reflector antennas (l 1m) • Hybrid antennas – Wire reflectors (l 1m) – Reflectors with dipole feeds 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 22 Fundamental antenna equations P(q,,n) A(q,,n )I(q,,n)n • Effective collecting area A(qn) m2 • On-axis response Ao=hA • Normalized pattern (primary beam) A(qn) = A(qn)/Ao • Beam solid angle A= ∫∫ A(n,q,) d all sky 2 • AoA = l 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 23 Aperture-Beam Fourier Transform Relationship • Aperture field distribution and far-field voltage pattern form FT relation f l g(u)e2iludu aperture • u = x/l • lsin q • Uniformly-illuminated aperture gives sinc function response qD P(q) sinc2 l • qHPBW l/D 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 24 Antenna efficiency On axis response: A0 = hA Efficiency: h = hsf . hbl hs ht hmisc rms error s hsf = Reflector surface efficiency Due to imperfections in reflector surface 2 hsf = exp[(4s/l) ] e.g., s = l/16 , hsf = 0.5 hbl = Blockage efficiency Caused by subreflector and its support structure hs = Feed spillover efficiency Fraction of power radiated by feed intercepted by subreflector ht = Feed illumination efficiency Outer parts of reflector illuminated at lower level than inner part hmisc= Reflector diffraction, feed position phase errors, feed match and loss 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 25 Interferometry • Largest fully-steerable dishes have D~100m – l/D ≤ 10-4 at cm wavelengths • Sub-arcsecond resolution impossible • Confusion limits sensitivity – Collecting area limited to D2/4 • Tracking accuracy limited to s~1” – Gravitational deformation – Differential heating – Wind torques – Must keep s < l/10D for accurate imaging/photometry • Interferometers consisting of multiple small dishes mitigate these problems 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 26 Interferometry • Interferometry = Aperture Synthesis – Combine signals from multiple small apertures – Technique developed in the 1950s in England and Australia. – Martin Ryle (University of Cambridge) earned a Nobel Prize for his contributions. • Basic operation: Correlation = multiply and average separate voltage outputs 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 27 Interferometry • We seek a relation between the characteristics of the product of the voltages from two separated antennas and the distribution of the brightness of the originating source emission. • To establish the basic relations, consider first the simplest interferometer: – Fixed in space – no rotation or motion – Quasi-monochromatic – Single frequency throughout – no frequency conversions – Single polarization (say, RCP). – No propagation distortions (no ionosphere, atmosphere …) – Idealized electronics (perfectly linear, perfectly uniform in frequency and direction, perfectly identical for both elements, no added noise, …) 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 28 The two-element interferometer • Radiation reaching 1 is delayed relative to 2 • Output voltage of antenna 1 lags by geometric delay tg = b.s/c • Multiply outputs: 2 V1V2 V cost cost t g V 2 cos2t t cost 2 g g • Time-average the result V 2 R VV cost 1 2 2 g multiply average 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 29 The two-element interferometer V 2 R VV cost 1 2 2 g • Sinusoidal variation with changing source direction: fringes bcosq t g c d bsinq 2 dq l • Fringe period l/bsinq • Phase measures source position (for large b) 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 30 The two-element interferometer V 2 R VV cost 1 2 2 g • Sensitivity to a single Fourier component of sky brightness, of period l/bsinq • Antennas are directional • Multiply correlator output by antenna response (primary beam) 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 31 Adding more elements • Treat as N(N-1)/2 independent two-element interferometers • Average responses over all pairs • Synthesized beam approaches a Gaussian • For unchanging sources, move antennas to make extra measurements 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 32 Extended sources V 2 • Point source response of our cosine correlator: R VV cost 1 2 2 g V 2 S A A 1/ 2 2 n 1 2 • Treat extended sources as the sum of individual point sources – Sum responses of each antenna over entire sky, multiply, then average R V d V d c 1 1 2 2 * • If emission is spatially incoherent n R1 n R2 0 for R1 R2 • Then we can exchange the integrals and the time averaging b.s Rc In scos2n d c 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 33 Odd and even brightness distributions b.s • Response of our cosine correlator: Rc In scos2n d c • This is only sensitive to the even (inversion-symmetric) part of the source brightness distribution I = IE+IO • Add a second correlator following a 90o phase shift to the output of one antenna V 2 R VV sint s 1 2 2 g b.s Rs In ssin2n d c I IE IO = + 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 34 Complex correlators A = (R 2+R 2)1/2 • Define complex visibility V = R -iR = Aei c s c s = tan-1(R /R ) • Thus response to extended source is s c b.s Vn In sexp 2i d l • Under certain circumstances, this is a Fourier transform, giving us a well established way to recover I(s) from V – 1) Confine measurements to a plane – 2) Limit radiation to a small patch of sky • Use an interferometer to measure the spatial coherence function Vn and invert to measure the sky brightness distribution 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 35 Digression: co-ordinate systems • Baseline b measured in wavelengths: the (u,v,w) co-ordinate system – w points along reference direction so – (u,v) point east and north in plane normal to w-axis – (u,v,w) are the components of b/l along these directions • In this co-ordinate system, s = (l,m,n) – l,m,n are direction cosines – n = cos q = (1-l2-m2)1/2 b.s V I sexp 2i d n n l dldm d I (l,m) 1l 2 m2 V u,v, w n exp 2iul vm wndldm n 2 2 1l m 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 36 Measurements confined to a plane • All baselines b lie in a plane, so b = r1-r2 = l(u,v,0) • Components of s are then l,m, 1l 2 m2 e2i(ulvm) V u,v, w 0 I (l,m) dldm n n 2 2 1l m • This is a Fourier-transform relationship between Vn and In, which we know how to invert 2 2 • Earth-rotation synthesis (so is Earth‟s rotation axis): 1l m cosq sec – E-W interferometers – ATCA, WSRT • VLA snapshots 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 37 All sources in a small region of sky • Radiation from a small part of celestial sphere: s = so+s • n = cos q = 1 – (q2/2) I (l,m) V u,v, w e2iw n exp 2iul vm wq 2 / 2dldm n 2 2 1l m • If wq 2 << 1, i.e. q~(l/b)1/2 then we can ignore last term in brackets I (l,m) Vu,v, w e2iw n e2i(ulvm)dldm n 2 2 1l m 2iw • Let Vn u,v,w e Vn(u,v,w) I (l,m) V u,v n e2i(ulvm)dldm n 2 2 1l m • Again, we have recovered a Fourier transform relation • Here the endpoints of the vectors s lie in a plane • Break up larger fields into multiple facets each satisfying wq 2 << 1 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 38 All sources in a small region of sky • This assumption is valid since antennas are directional V u,v A (l,m)I (l,m)e2i(ulvm)dldm n n n • Primary beam of interferometer elements An gives sensitivity as a function of direction • Falls rapidly to 0 away from pointing centre so • Holds where field of view is not too large – centimetre-wavelength interferometers • EVLA • MERLIN • VLBA • EVN • GMRT 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 39 Visibility functions • Brightness and Visibility are Fourier pairs. • Some simple and illustrative examples make use of „delta functions‟ – sources of infinitely small extent, but finite total flux. 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 40 Visibility functions • Top row: 1-dimensional even brightness distributions. • Bottom row: The corresponding real, even, visibility functions. Inq Vn(u) 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 41 Inverting the Fourier transform • Fourier transform relation between measured visibilities and sky brightness I (l,m) V u,v n e2i(ulvm)dldm n 2 2 1l m • Invert FT relation to recover sky brightness I l,m n V (u,v)e2i(ulvm)dudv 2 2 n 1l m • This assumes complete sampling of the (u,v) plane • Introduce sampling function S(u,v) I D l,m n V (u,v)S (u,v)e2i(ulvm)dudv 2 2 n n 1l m D • Must deconvolve to recover In : In In B • See Katherine Blundell‟s lecture for details 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 42 Seems simple? • We have discovered a beautiful, invertible relation between the visibility measured by an interferometer and the sky brightness • BUT: • Between the source and the correlator lie: – Atmosphere – Man-made interference (RFI) – Instrumentation • The visibility function we measure is not the true visibility function • Therefore we rely on: – Editing: remove corrupted visibilities – Calibration: remove the effects of the atmosphere and instrumentation obs true * true Vij Gij Vij GiGj Vij 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 43 What about bandwidth? n • Practical instruments are not monochromatic G • Assume: – constant source brightness – constant interferometer n0 n response over the finite bandwidth, then: 1 n c n / 2 Vn n In (s)exp 2intg dv d n c n / 2 • The integral of a rectangle function is a sinc function: V I (s)sinc(nt )exp(2in t )d n n g c g • Fringe amplitude is reduced by a sinc function envelope • Attenuation is small for ntg<< 1, i.e. nq qsn 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 44 What about bandwidth? • For a square bandpass, the bandwidth attenuation reaches a null at an angle equal to the fringe separation divided by the fractional bandwidth: nn0 • Depends only on baseline and bandwidth, not frequency Fringe Attenuation function: b n sinc q l n c sin q bn 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 45 Delay tracking • Fringe amplitude reduced by sinc function envelope • Eliminate the attenuation in any one direction so by adding extra delay t0 = tg • Shifts the fringe attenuation function across by sin q = ct0/b to centre on source of interest • As Earth rotates, adjust t -1 continuously so | t0 - tg| < (n • nq qsn • Away from this, bandwidth smearing: radial broadening • Convolve brightness with rectangle nqn 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 46 Field of view • Bandwidth smearing – Field of view limited by bandwidth smearing to q qsnn Source – Split bandwidth into many narrow channels to widen e FOV q • Time smearing NCP – Earth rotation should not move source by 1 synthesized l/D beam in correlator averaging time – Tangential broadening – q Pq 2t where Interferometer Primary Beam s Fringe Separation Half Power P=86164s (sidereal day) l/b 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 47 What about polarization? • All the above was derived for a scalar electric field (single polarization) • Electromagnetic radiation is a vector phenomenon: – EM waves are intrinsically polarized – monochromatic waves are fully polarized • Polarization state of radiation can tell us about: – the origin of the radiation • intrinsic polarization, orientation of generating B-field – the medium through which it traverses • propagation and scattering effects – unfortunately, also about the purity of our optics • you may be forced to observe polarization even if you do not want to! 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 48 The Polarization Ellipse • From Maxwell‟s equations E•B=0 (E and B perpendicular) – By convention, we consider the time behavior of the E-field in a fixed perpendicular plane, from the point of view of the receiver. k E 0 transverse wave • For a monochromatic wave of frequency n, we write Ex Ax cos2nt x Ey Ay cos2nt y – These two equations describe an ellipse in the (x-y) plane. • The ellipse is described fully by three parameters: – AX, AY, and the phase difference, = Y-X. 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 Elliptically Polarized Monochromatic Wave The simplest description of wave polarization is in a Cartesian coordinate frame. In general, three parameters are needed to describe the ellipse. If the E-vector is rotating: – clockwise, wave is „Left Elliptically Polarized‟, – counterclockwise, is „Right Elliptically Polarized‟. equivalent to 2 independent Ex and Ey oscillators 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 Polarization Ellipse Ellipticity and P.A. • A more natural description is in a frame (xh), rotated so the x-axis lies along the major axis of the ellipse. • The three parameters of the ellipse are then: Ah : the major axis length tan c AxAh : the axial ratio : the major axis p.a. • The ellipticity c is signed: c > 0 REP c < 0 LEP c = 0,90° Linear (=0°,180°) c = ±45° Circular (=±90°) 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 Circular Basis • We can decompose the E-field into a circular basis, rather than a (linear) Cartesian one: E AReR ALeL – where AR and AL are the amplitudes of two counter-rotating unit vectors, eR (rotating counter-clockwise), and eL (clockwise) – NOTE: R,L are obtained from X,Y by a phase shift • In terms of the linear basis vectors: 1 A A2 A2 2A A sin R 2 X Y X Y XY 1 A A2 A2 2A A sin L 2 X Y X Y XY 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 Circular Basis Example • The black ellipse can be decomposed into an x- component of amplitude 2, and a y-component of amplitude 1 which lags by ¼ turn. • It can alternatively be decomposed into a counterclockwise rotating vector of length 1.5 (red), and a clockwise rotating vector of length 0.5 (blue). 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 Stokes parameters • Stokes parameters I,Q,U,V – defined by George Stokes (1852) – scalar quantities independent of basis (XY, RL, etc.) – units of power (flux density when calibrated) – form complete description of wave polarization – Total intensity, 2 linear polarizations and a circular polarization 2 2 2 2 • I = EX + EY = ER + EL 2 2 • Q = I cos 2c cos 2y = EX - EY = 2 ER EL cos RL • U = I cos 2c sin 2y = 2 EX EY cos XY = 2 ER EL sin RL 2 2 • V = I sin 2c = 2 EX EY sin XY = ER - EL • Only 3 independent parameters: – I2 = Q2 + U2 + V2 for fully-polarized emission – If not fully polarized, I2 > Q2 + U2 + V2 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 Stokes parameters • Linearly Polarized Radiation: V = 0 – Linearly polarized flux: P Q2 U 2 – Q and U define the linear polarization position angle: tan2 U /Q – Signs of Q and U: Q > 0 U > 0 U < 0 Q < 0 Q < 0 U > 0 U < 0 Q > 0 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 Polarization example • VLA @ 8.4 GHz – Laing (1996) • Synchrotron radiation – relativistic plasma – jet from central “engine” – from pc to kpc scales – feeding >10kpc “lobes” • E-vectors – along core of jet – radial to jet at edge 3 kpc 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 Sensitivity • The radiometer equation for point source sensitivity: 2kBTsys s S Aeff N(N 1)tintn • Increase the sensitivity by: • Increasing the bandwidth • Increasing the integration time • Increasing the number of antennas • Increasing the size/efficiency of the antennas • Surface brightness sensitivity is MUCH worse • synthesized beam << beam size of single dish of same effective area 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 57 Summary: why should I use radio telescopes? • Unparalleled resolution – Astrometry: positions, velocities, distances – High-resolution imaging • Unique probe – Radio waves not scattered by dust: probe inner Galaxy – Low-energy, long-lived non-thermal emission – Spectral lines probe physical conditions of gas • Signature of high-energy processes; link to X-ray emission • Exotic phenomena – Coherent emission (pulsar phenomenon) • New “Golden Age” – New telescopes opening up new parameter space 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 58 Radio Astronomy: A Golden Age! WSRT EVLA EVN LOFAR GMRT VLBA ATCA SKA? ALMA MeerKAT LBA ASKAP 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 59 Equations to take away • Resolution of an interferometer l q bmax • Field of view of an interferometer l q D • Sensitivity of an interferometer 2kBTsys s S Aeff N(N 1)tintn • Relation between visibility and sky brightness I (l,m) V u,v n e2i(ulvm)dldm n 2 2 1l m 2nd School on Multiwavelength Astronomy, Amsterdam , June 2010 60